Moral hazard is the additional health care that is consumed because of insurance. Conventional theory regards moral hazard as being welfare decreasing, but this does not account for the additional health care that is purchased because of the transfer of income from those who purchase insurance and remain healthy, to those who purchase insurance and become ill. Recognition of this income effect suggests that a portion is welfare decreasing but another portion is welfare increasing. The welfare-increasing portion can be estimated by the additional health care that an insured consumer would have purchased if, instead of paying for any care, the insurer had paid him a cashier's check in an amount equal to the portion of his or her health care expenditures covered by the insurer under the standard insurance policy. Rather than conducting an expensive RAND-type of health insurance experiment, this study proposes to estimate the welfare-increasing portion of the additional health care due to insurance using the 1996-2008 Medical Expenditure Panel Survey, and to do so for (a) the general population of Americans, (b) those Americans with 9 of the 15 priority health conditions separately, and (c) all 15 priority health conditions combined. It will use a 7-step process to (1) estimate health care spending using an instrumental variables approach to correct for the endogeneity of private and public health insurance, taking into account the possible choice of public insurance, (2) for each uninsured, estimate the income elasticity of demand from the income coefficients in step 1, (3) for the privately insured, estimate the portion of total spending that would have been paid for by the insurer (excluding cost sharing), (4) for each uninsured, use the coefficients from step 3 to estimate the amount of spending paid for by the insurer if they had been insured, (5) for each uninsured, use the income coefficients from step 2 to estimate the change in uninsured health care spending if they were insured with a contingent claims insurance policy that paid off with the same spending as in step 4, (6) for each uninsured, use the coefficients from step 1 to estimate the moral hazard spending if they had been insured, and (7) for each uninsured, the change in health care spending had the individual been insured by a contingent claims contract from step 5 is multiplied by the estimated spending in the uninsured state to determine the efficient portion of the additional health care spending. This efficient spending with a contingent claims policy is then compared to estimated overall moral hazard spending from step 6 to determine the amount of inefficient moral hazard. An analogous procedure will be repeated for the insured and for the entire sample. This study represents the first to estimate the portion of moral hazard that is efficient. This information wil be used to make inferences about the proportion of the total health care spending in the U.S. that represents efficient moral hazard, and also will be used to evaluate appropriateness of imposing cost sharing on insured persons with the various priority health conditions.